The objective of this study was to evaluate the effect of population density on mortality after traffic crashes. Subjects were the drivers of vehicles requiring towing after collisions, sampled by NASS-GES during 1994-1998. Cases were classified by population of crash location (greater or lesser than 25,000) and by population density of the driver's county of residence (using ZIP code and Census data). Cases were also categorized by age, belt use and vehicle speed. Analytic methods for weighted, stratified survey data were used to estimate effects on mortality. A total of 190,721 cases with a specified crash location were identified in the sample, of which 93.7% had a specified population at the crash location, and 94.1% could be linked to the driver's county data. Mortality was higher in locations with populations less than 25,000 and was inversely proportional to the driver's county population density. Age, belt use, and vehicle speed also affected mortality significantly, but the effect of rural location remained after controlling for these factors (O.R. 2.10, 95% C.I. 1.62-2.73). The excess risk for residents of rural areas to die in traffic crashes can be attributed in part to post-crash factors.